1. Field of the Invention
The present invention generally pertains to processing and comparing data in a data store like a relational database or full text search engine. More specifically, a method for converting arbitrary strings to numeric representations is disclosed so that complex comparisons can be carried out on the numeric data values.
2. The Relevant Technology
Group collaboration poses challenges for identifying key documents and records in a database. Each group may construct a unique way to identify the document with their own naming convention and guidelines. A purchase order number to the customer, a bill of sale identifier to the vendor, and a packing list to the logistics provider could represent a single document or record. Although the problem of representing a single record with multiple identifiers has already been solved, it is not trivial to provide a generic way to use these various identifiers within a multi purpose data store environment. Conventionally, if the identifier is composed of a string and numeric value, separate methods would be required to handle each component (i.e., PO-12345 could be separated into “PO” and “12345”). If the logistics provider used an identification system that includes different numeric and string components (i.e. EFG-123-123-45-MN), the data store would have to be modified separately to handle the new scheme (“EFG” “123” “123” “45” “MN”). For each new participant in the collaboration process, new schemes or strict rules would be required to allow participation using their own identification system. This problem is made more difficult if the data store has primitive data types without custom data types.
Hence, there is a need for converting arbitrary string data into a useful format for complex comparisons. The method should allow for the greatest variety of user provided schemes and map them to the simplest of data types—the numeric field.